DGRNet: Disagreement-Guided Refinement for Uncertainty-Aware Brain Tumor Segmentation

2026-03-22

AI Summary

This work addresses the lack of reliable single-model uncertainty quantification in brain tumor MRI segmentation and the underutilization of semantic information from radiology reports. To this end, the authors propose DGRNet, a novel framework that estimates segmentation uncertainty through a multi-view disagreement mechanism built upon a shared encoder-decoder architecture augmented with four lightweight view adapters. For the first time, a text-conditioned refinement module is introduced to semantically guide corrections in high-uncertainty regions using clinical reports. To prevent view collapse, a diversity-preserving training strategy is devised, combining pairwise similarity penalties with gradient isolation. Evaluated on TextBraTS, DGRNet achieves a 2.4% improvement in Dice score and an 11% reduction in HD95, significantly outperforming existing methods while providing clinically meaningful uncertainty estimates.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Segmentation
• Reasoning under Uncertainty: Uncertainty Representations

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Accurate brain tumor segmentation from MRI scans is critical for diagnosis and treatment planning. Despite the strong performance of recent deep learning approaches, two fundamental limitations remain: (1) the lack of reliable uncertainty quantification in single-model predictions, which is essential for clinical deployment because the level of uncertainty may impact treatment decision-making, and (2) the under-utilization of rich information in radiology reports that can guide segmentation in ambiguous regions. In this paper, we propose the Disagreement-Guided Refinement Network (DGRNet), a novel framework that addresses both limitations through multi-view disagreement-based uncertainty estimation and text-conditioned refinement. DGRNet generates diverse predictions via four lightweight view-specific adapters attached to a shared encoder-decoder, enabling efficient uncertainty quantification within a single forward pass. Afterward, we build disagreement maps to identify regions of high segmentation uncertainty, which are then selectively refined according to clinical reports. Moreover, we introduce a diversity-preserving training strategy that combines pairwise similarity penalties and gradient isolation to prevent view collapse. The experimental results on the TextBraTS dataset show that DGRNet favorably improves state-of-the-art segmentation accuracy by 2.4% and 11% in main metrics Dice and HD95, respectively, while providing meaningful uncertainty estimates.

Problem

• uncertainty quantification
• brain tumor segmentation
• radiology reports
• ambiguous regions
• clinical decision-making

Innovation

• uncertainty-aware segmentation
• multi-view disagreement
• text-conditioned refinement
• diversity-preserving training
• brain tumor segmentation